Riot’s Preseason Q&A on the forums has now concluded. Follow the links below for all the responses from the balance team:
We’re about wrapping up now, we’ll have a finalized rundown later this week, but thanks for participating! We’ll probably close up comments by end of tonight.
Welcome to the preseason Q&A, where myself and the design team will be fielding major questions you might have about the current preseason. Before we get to that, I wanted to write up some early first impressions we’ve had so far. There are certainly some outliers we want to address before we can get a deeper understanding, but initial thoughts might quell some concerns. A few weeks from now we’ll have a more high-level overview.
In the meantime…
Preseason so far
It’ll take a while for things to shake down, but I’ll dig into some first thoughts. If, after reading this, there’s anything not mentioned you’d like to get our thoughts on, we’ll be answering questions between 1pm-3pm (with some less regular posting throughout the rest of the day too (where time permits) between other work).
General Game Elements
Overall, we’re pretty happy with how the new masteries are going, with some exceptions. The first of those is that there are some that are definitely too strong. The most notable of those is Warlord’s Bloodlust, and we’ll be hotfixing that today to work on champions only, rather than against all targets. Might also need to make some changes so that this mastery’s just not just dramatically more effective on Trynd/Yasuo than other champs (or make changes to Trynd/Yasuo if needs be). Other likely outliers are Deathfire Touch (probably too strong on DoT champs, potentially too weak on others) and some of the Resolve tree on health stacking tanks.
On the other hand there are also some champions that aren’t sufficiently well served by the current masteries. Starting soon (potentially patch 5.24) we’ll start adding a few extra masteries in to flesh out the trees a bit more, probably beginning with the Tier 1 Cunning masteries, since that juncture currently doesn’t feel like it offers many champs a good choice.
- Minion Blocking
We made some under the hood changes to make minion blocking more predictable and less random in 5.22, but it looks like we also made minions feel too large. We’ll be looking to adjust this to make it feel “right” again, starting with some tuning to how minion blocking’s calculated in 5.23.
- The Jungle
We’re seeing certain champions struggle more than they should at present. We’ve got some buffs to Hunter’s Talisman and Refillable Potion coming as a result to help out some early clears. We’re also seeing some magic damage dealing junglers in particular have issues, so are lowering Magic Resistance on some monsters to help them specifically. If you’re currently struggling in the jungle, you may want to try Krugs (or Gromp) > double buffs, as that seems to be a fairly stable approach.
As far as Rift Herald goes, we’re seeing it do some cool stuff so far and it seems to be generating the sort of teamwork and reward for controlling the top part of the river we’d hoped. That said, the Herald is looking like it’s a bit too punishing, so we’re reducing its damage output somewhat while trying to keep it tanky. Hopefully this reduces the need to recall after taking it for many champions.
We’re currently seeing turrets die too easily. Some of that’s assumedly due to the reduction in effective health tower health in 5.22, some of it’s also going to be due to a bug that’s removing their intended growth in Armor/MR over time. We’ll be fixing that bug in 5.23, and increasing outer tower health a bit, then looking at whether towers are still falling too fast or not.
Graves is significantly too strong, Kog’maw definitely too weak. We’ll be nerfing and buffing those two respectively in today’s hotfix, with Graves losing some burst and early harass power from Q (damage and mana cost nerfs) and a bit of base MS. Kog by contrast’s getting a little bit of early AD and a better base attack speed. In both cases we suspect we’ll need to make further adjustments as well. We’d rather chip at some of the problem quickly and then reassess than try and do everything at once without sufficient observation time.
Current thinking is that a few other marksmen probably also need adjustments, though just in the form of regular balance changes rather than a hotfix. For 5.23 we’re looking at cutting a bit of power from Lucian and MF, both of whom seem too dominant, and we’ll be fixing a couple of small bugs on Ashe’s Q that should help her out a bit. We’re also looking at Quinn to try and shift some of her power from raw damage back into utility, though aren’t sure yet whether we’ll have changes we’re happy with for 5.23 or not.
Physical damage assassins seem to be in a rough spot after the removal of Brutalizer and changes to Last Whisper. In the short term, we’re looking to give some of them a bit of love through buffs to their kits (Kha’zix/Zed), longer term we’ll also want to find some other solutions as well (potentially assassin focused items, or item changes).
- Other Champion Changes
It’s not preseason specific, but we’ve also got some significant follow up work on Mordekaiser coming soon. Statikk should be covering that in a separate thread in the Dev Corner within the next couple of days, with some context on what we felt worked and what didn’t from the Morde rework in 5.16, plus thoughts on what our next steps are and why when it comes to addressing issues Morde has.
The next patch will also include a general sizing pass on a large number of champions, with model sizes of around 40 champions changes by between -10% and +10%. The goal there’s to align champion size better with each champion’s role (tanks should look bigger and tougher, squishies like ADCs smaller and more…squishy) while at the same time removing some visual discrepancies (Lucian’s not intended be 7 foot plus tall for example).
Statikk Shiv’s looking too generically strong. We’ll be hitting its damage as a result, while increasing its damage against minions to try and focus it more as a waveclear item.
Rageblade is just all-around way too strong. As a first step we’ll be increasing its cost a few hundred gold, but I imagine we’ll need to look at further power cuts at some point as well.
- Clean Up, Bug Fixes
We’ve got a number of start of game things to sync up (bunch of stuff that needs to be shifted forwards up to 15s to match the earlier game start, like Bard Chime spawns, Skarner Crystals, Smite charges etc), plus a number of bug fixes to things that changed in 5.22.
- Game Duration
Game duration, and snowballiness, has been one of the things we’ve spent the most time looking at and talking about since the preseason launch. So far it’s looking like the average game is a couple of minutes faster. There are some changes coming in 5.23 that will likely impact that, in particular a change to Homeguard that gives a modified version to every player after 20 minutes, somewhat tougher turrets (see above), and changes to minion strength that cause them to push more if their team is winning (giving winning teams opportunities to siege/losing teams opportunities to farm and catch up).
In terms of direction, two of the big things we’re looking at are the concerns that games are ending too often before late game and that a single mistake mid game may be deciding things too much, too often. We’re also looking at whether games are too stompy in general. Some increase in stompiness is expected for the first patch of the preseason, as people experiment with new things, that’s something that should fall off fairly quickly though. If not, we’ll need to make some follow up changes.
Yes, we’ve seen a noticeable increase in surrendering around the 20 minute mark post 5.22. That’s pretty standard after a major patch like this, given some champions tend to be out of line and people are experimenting with new stuff, so games can feel decided pretty early on. It’s something we’d expect to drop off noticeably over the next couple of patches though as we address major issues and people figure out what works and what doesn’t post changes. If that doesn’t happen we’ll need to dig into what else is going on.
We’ve just started talking about which class (or sub class) to look at next. We’re currently trending towards immobile mages. Given the need for preseason followup though it’ll also be a little while before the team’s ready to start heavy work on whichever class we go with.
Yeah, we’re also seeing Malphite and Rammus, other champions who are rewarded for stacking armor, do well right now given the number of AD damage dealers being played. I don’t think we’ll make any changes to those champs individually until we’ve got the bigger picture balanced, it seems reasonable to believe that they’re just reflecting a temporary meta state rather than themselves generally overpowered.
There were 4 teams responsible for the in game changes:
Game Systems – things like Masteries, Items, minions etc
Live Gameplay – Rift Herald, ongoing balance tuning
Champion Update – Marksman
Playtest – Sanity checks on content before it goes Live
It’s really a bigger team effort though, and the collaboration between teams is pretty huge.
Also worth calling out that we had help from individuals on various other teams as well. Some of the champion team helped out with masteries, couple of folks from the team that worked on the Bilgewater event helped with Rift Herald’s model etc.
Our goal with preseason was to keep games around the same average length as they currently were (around 33 minutes on average). Games in 5.22 look like they’re a couple of minutes shorter so far, at around 31 minutes, that’s likely to change though in 5.23 as we get in changes that either missed the cutoff for 5.22 or were bugged (minion changes, tower scaling defense respectively).
Hey man! We’re going to be watching, reading, listening, and playing a lot over the next few weeks to get a feel for where the new lay of the land is. Right now win rates are still being highly impacted by people learning new optimized item builds, masteries, abilities etc. Vayne as an example has an~8% delta in win rate between her most frequent build and her highest common win rate build (referencing champion.gg in this case since that’s available to everyone). If you look at the why, it’s because a subset players are still rushing BORK, which looks like it’s proving to no longer be the optimal build path. Just pulling that as one example. Over the next few weeks, as players discover those optimal builds, I’m expecting we’ll see the standard deviation for individual champ win rates get smaller, time will definitely tell though!
I know you referenced Kha’zix in particular, and he’s definitely taken a hit this preseason (to me it feels like it’s actually more about the removal of Trailblazer as opposed to Tiamat passive, but that’s just my $0.02). We’re investigating how we can give him a bit of love, but want to wait for things to settle before getting more aggressive with changes. Also note I used to play Kha’zix a fair bit so I feel your pain :).
Hope this at least touches on some of your concerns
Yes! I’ve given quite a bit of thought to this actually. One of the things that I think is important to get right here is how to balance what feels good for Quinn versus what feels fair for the opponent. Though I don’t have specific changes for 5.23 to make on this front, I’m keeping an eye on Quinn (her ult especially) to see what kind of adjustments need to be made over the course of the next few patches.
What I’m currently feeling is that it’s too easy/annoying to get knocked out of R or the channel (I think minions cancelling the channel, for example, may be out of hand), and that a cooldown or different restriction on the R may be appropriate in the long term in mitigating some of those more earflicky interactions associated with her R at the moment.
I think on the Quinn update generally, we actually didn’t see her as “heavily reworked.” What we intended for her was primarily the changes to her ultimate, and then supporting changes to make that work well. Admittedly, we missed the mark on the proper supporting changes, hence our reaction to re-institute a utility-focused Q in 5.23, but we do think that the changes to her ultimate open up a lot of doors for her both strategically and tactically. That said, I know there are old Quinn mains that won’t agree with that.
It’s certainly a possibility that we’ll be looking out for. If she ever becomes popular, players will be quick to pick it up and let us know.
Following that conversation, I made some adjustments to where E places you when you cast from max range in hopes to address this issue a bit. I don’t currently think her chase is too bad at the moment, but that isn’t to say I couldn’t be convinced otherwise.
This was done on purpose. Caitlyn is a (perhaps the) premiere siege marksman. She should feel good about attacking towers. We added this mechanic back in on Caitlyn because it made sense that she wouldn’t feel a tension between shooting towers and minions.
I’ll go into some elements of our approach. As an individual contributor to the marksmen updates, it will only reflect my perspective, but hopefully it will be helpful:
1) Winrate data is heavily skewed at the moment due to three factors.
First, a couple of champions are way too strong (e.g., Graves). They will naturally lower the win rate of all other marksmen.
Second, champions with new mechanics ought to go down in win rate as people learn to use them. This includes itemization changes (e.g., we are confused as to why players aren’t trying Essence Reaver on Sivir).
Third, not every champion will have a 50% win rate at balance. For example, as we differentiate marksmen, you’ll have to pick them in the right circumstances – If you pick Vayne on a team that desperately needs wave clear, you’ll probably be worse off than if you’d picked Sivir. Champs that have well defined niches will tend to be more powerful than their win rate suggests.
2) We aren’t sure where every champion sits. Obviously we have our own internal assessments of champion power, but they actually differ markedly from a lot of players’ perspectives. Unless we take the time to carefully watch and play the champions on Live for a while, we risk just playing musical chairs, shifting a new set of marksmen from apparently weak to actually overpowered. Further, we can’t just balance champs. We need to listen to you, the player base, about how you want to use champs in order to get the _right _changes in place. Technically, we could just balance any champ just through changing their base HP and HP per level, but it wouldn’t lead to a satisfying experience.
3) That is all very nebulous. So here are some specific answers.
We’re hotfixing Graves and Kog’Maw today (nerf and buff respectively). I’d look some obvious changes in 5.23 and a substantial amount of fine tuning in 5.24.
Caitlyn’s Q is indeed too weak. Even if we had to nerf other elements of her, it would be worth buffing the ability to let her actually reach towers and make skillfully avoiding it matter when playing against. I believe duo bot Twitch is very powerful and that players who are routinely failing with him are probably not good at the champion or are fixated on a strategy that is now less effective (teamfight AoE) as opposed to opportunistic assassination. Even the data supports this. His win rate is almost 50% when played in his best role. He’s not doing well in the jungle, and we’ll have to assess how much we can support him there.
* I don’t have a good sense for Sivir and Ashe’s balance right now, so perhaps someone else can reply to that. Anecdotally, they feel effective given their utility.
Having played a decent bit of Kha’zix both in normals and ranked over the years, he’s definitely someone that was hurt more than we’d like due to the preseason changes (looking at you Trailblazer). While we want to avoid any heavy handedness right now due to the number of things that have changed, there’s definitely room to give him a bit of love, especially in regards to his early game.
I played him last night in order to get a better feel on Live, and after Twitch accidentally took my large Krug early which kinda boned me (hit lvl 2 at 3:55), I still managed to do alright. Like I said though, definitely want to show him some love in the near future, as well as potentially look at something more long term.
Hope that helps!
We see GW as a useful safety valve on healing/high sustain champions, and one that’s probably best accessed in game in response to an enemy getting out of control (e.g. a really fed Swain), rather than accessed via champ select before you know how needed it’ll be. GW on individual kits has the potential to lead to hard counter situations, and while we do want to reward good choices in champ select we’d like to avoid games that feel predecided before the loading screen.
We don’t have any plans to remove blind from Teemo and I’d imagine will at some point look to use it on at least one other champion. Would want to improve the visual/audio feedback for the blinded player to help them play around the effect before doing so though. The important thing would be putting it on a champion where it doesn’t leave the enemy without sufficient means to respond. Teemo’s really range limited, doesn’t burst people down generally and lacks other survivability tools when in the thick of the fight. Quinn by contrast had much longer range on her blind, has a powerful mid fight reposition and is adept at forcing fights with targets of her choice.
I’m excited that the changes for 5.22 did not seem to make him too powerful, and given that, I’d like to keep pushing that direction in an upcoming patch if there’s an opportunity. An example of something we tested internally but didn’t end up shipping (for power concerns) was that the reset on Ambush had more Movement Speed and dropped into stealth instantly instead of with the 1.5 second fade time. This isn’t necessarily exactly what we’d do with him moving forward, but making that added mechanic feel right is probably the basic idea.
I know Jag had a very specific reason as to why he kept Herald a one-person buff, so I’ll poke him to answer here.
For the CDR topic, I wonder if the ideal state of Mages is that they are guaranteed to hit CDR cap without much tradeoff or if they can be balanced in a way that investing into CDR is a meaningful decision (ie: giving up MR and maybe even Sorc Boots to double down on cooldowns for the additional utility). A world where a mage decides between high-damage, higher-cooldown abilities vs. lower-damage, utility-focused, lower-cooldown abilities seems like an interesting decision if done well.
In other words, right now there’s a large mentality shift of “I will always hit the CDR cap” to “I need to think try very hard to hit the CDR cap.” Additionally, since outside of the keystone masteries there hasn’t been anything more powerful injected into Mages, it might feel like a power decrease but we have a lot of noise and data we’ll need to parse through to see if this is actually true. Are mages weak? Are mages properly calibrated against their core stats (CDR, AP, mana regen)? Honestly speaking I want to see more before just saying “if you raise gold costs on AP and make it more difficult to hit the CDR cap, of course mages are going to be weak,” – it’s hard to assess that without seeing how the ecosystem of the game actually plays out, not to mention the strength of Deathfire Touch.
Additional note on other champions getting more ‘free’ access to CDR than mages. These kind of comparisons are really weird to me, because CDR is one of those stats that’s very closely tied to the balance of the kit using it. CDR on marksmen tends to be very singular in purpose – damage – and often dependent on proper positioning to use well. Most Mage abilities tend to be fight-starters or fight-stoppers, meaning they have the ability to dramatically change the landscape or flow of a fight on a single ability cast. CDR to them is far more valuable and, in their natural item builds, needs to be ‘priced’ accordingly.
If stats were just generically priced without accounting for the stat affinities of the classes purchasing them, I think we’d get some really, really weird scenarios.
I edited in an answer regarding that, but I do feel like we’re in a bit of a vacuum comparison. A mage using abilities on cooldown can often stop or outright deny a champion threat. A marksman can do the same via a steady stream of damage but we’re basically talking variance in power. Mages have high variance – high highs, low lows (no abilities to cast), while Marksmen have less highs but more consistent output. In a balanced world, these two types of damage output would be well balanced against each other, but Mages also get the addition of utility, which needs to be considered as well.
I suppose this is a long way of saying damage isn’t the only thing to compare and Mages vs. Marksmen can’t really be boiled down to individual stat comparisons and who gets access to what. A reverse example (not totally comparable) might be if Marksmen were demanding total AD scaling on Trinity Force procs, similar to Lich Bane.
I think I’m saying that Mages have access to a lot more breadth of power than Marksmen and their itemization currently reflects that as well. If a Mage were to become a magic-focused Marksman I think it’d be easier to balance them, but then you’d be effectively choosing between the two. All a marksman has – all that they are identified by – is their pure damage with key positioning. A mage can be that, but they can also be the one who picks off fleeing champs, or controls the battlefield, or catches and blows up a stray squishy.
The question is also what they do with those abilities. AD Casters get access to CDR very easy, but they can’t transform that CDR into an ability that starts and ends a teamfight on a single cast. An Ahri charm at late game can potentially mean a squishy target is dead without using any ability. An Ezreal Q cast on cooldown does not.
Really, really important distinction.
We increased item costs generally in preseason while modifying gold income and item premiums. We have seen a lot of early feedback on AP mages feeling punished by that, from the looks of it so far though some are struggling and some aren’t however, so we’ll need to be more tactical about changes than just straight AP item buffs across the board (Heimer and jungle Nid look too weak for example, Brand and Swain too strong).
I expect we’ll look at AP items in the coming year. Immobile mages in particular are one of the classes we’re thinking of tackling next and some item adjustments would likely be part of that process as with Marksmen/Juggernauts.
We are. In particular it’s getting champions to the CDR cap too early in the game. We’ll be looking at some changes, probably for 5.24, to make some of that CDR come later in the game.
We’ll be fleshing out the Tier 1 and Tier 2 keystones in Cunning, which should help somewhat there. For some champions though Ferocity’s probably always going to be more appealing (DoT mages especially will be really strongly drawn to DFT). Assassins by contrast are meant to be well served by the Cunning tree, so if they’re not getting decent performance from it we might need to make the Tier 3 keystones a bit stronger for them.
It’s looking like it’ll probably need some form of buff yeah.
We’d like DFT to be the sustained damage option and Thunderlord’s the burst damage one. Both should have reasonable scaling though, hence the ratios – we’re not looking to make one of them a non item dependent option.
Fervor of Battle’s proc damage, so doesn’t work with lifesteal, can’t crit etc.
Quite soon (almost certainly next rework of any size to go out).
I’m not sure our designers who work on that are in this Q and A, so I’ll answer for them! I would be somewhat opposed to this on the grounds that ARAM doesn’t test a wide variety of skills and because the random mode is about encouraging you to branch out, not go deep on a small number of champions. However, I think it would be cool if there were ways to express your mastery of ARAM as such — for example, ones more related to the map or champion classes.
Thresh is meant to be a ranged champion. When we updated how we tag champions as melee or ranged he got missed for some reason (not sure why myself sorry). That’ll be fixed in the next patch.
If you have any questions, feel free to ask me atÂ @NoL_ChefoÂ or e-mail me at email@example.com.